[infinispan-dev] Adjusting memory settings in template

Mon Sep 25 07:15:48 EDT 2017

> On 25 Sep 2017, at 12:37, Sebastian Laskawiec <slaskawi at redhat.com> wrote:
> 
> 
> 
> On Mon, Sep 25, 2017 at 11:58 AM Galder Zamarreño <galder at redhat.com> wrote:
> 
> 
> > On 22 Sep 2017, at 17:58, Sebastian Laskawiec <slaskawi at redhat.com> wrote:
> >
> >
> >
> > On Fri, Sep 22, 2017 at 5:05 PM Sanne Grinovero <sanne at infinispan.org> wrote:
> > On 22 September 2017 at 13:49, Sebastian Laskawiec <slaskawi at redhat.com> wrote:
> > > It's very tricky...
> > >
> > > Memory is adjusted automatically to the container size [1] (of course you
> > > may override it by supplying Xmx or "-n" as parameters [2]). The safe limit
> > > is roughly Xmx=Xms=50% of container capacity (unless you do the off-heap,
> > > that you can squeeze Infinispan much, much more).
> > >
> > > Then there are Limits, Requests and QoS in Kubernetes [3][4]. We are in
> > > bustable memory category so if there is additional memory in the node, we'll
> > > get it. But if not, we won't go below 512 MB (and 500 mCPU).
> >
> > I hope that's a temporary choice of the work in process?
> >
> > Doesn't sound acceptable to address real world requirements..
> > Infinispan expects users to estimate how much memory they will need -
> > which is hard enough - and then we should at least be able to start a
> > cluster to address the specified need. Being able to rely on 512MB
> > only per node would require lots of nodes even for small data sets,
> > leading to extreme resource waste as each node would consume some non
> > negligible portion of memory just to run the thing.
> >
> > hmmm yeah - its finished.
> >
> > I'm not exactly sure where the problem is. Is it 512 MB RAM/500 mCPUs? Or setting 50% of container memory?
> >
> > If the former and you set nothing, you will get the worse QoS and Kubernetes will shut your container in first order whenever it gets out of resources (I really recommend reading [4] and watching [3]). If the latter, yeah I guess we can tune it a little with off-heap but, as my the latest tests showed, if you enable RocksDB Cache Store, allocating even 50% is too much (the container got killed by OOM Killer). That's probably the reason why setting MaxRAM JVM parameters sets Xmx to 25% (!!!) of MaxRAM value. So even setting it to 50% means that we take the risk...
> >
> > So TBH, I see no silver bullet here and I'm open for suggestions. IMO if you're really know what you're doing, you should set Xmx yourself (this will turn off setting Xmx automatically by the bootstrap script) and possibly set limits (and adjust requests) in your Deployment Configuration (if you set both requests and limits you will have the best QoS).
> 
> Try put it this way:
> 
> I've just started an Infinispan ephermeral instance and trying to load some data and it's running out of memory. What knobs/settings does the template offer to make sure I have a big enough Infinispan instance(s) to handle my data?
> 
> Unfortunately calculating the number of instances based on input (e.g. "I want to have 10 GB of space for my data, please calculate how many 1 GB instances I need to create and adjust my app") is something that can not be done with templates. Templates are pretty simple and they do not support any calculations. You will probably need an Ansible Service Broker or Service Broker SDK to do it.
> 
> So assuming you did the math on paper and you need 10 replicas, 1 GB each - just type oc edit dc/<your_app> and modify number of replicas and increase memory request. That should do the trick. Alternatively you can edit the ConfigMap and turn eviction on (but it really depends on your use case).
> 
> BTW, the number of replicas is a parameter in template [1]. I can also expose memory request if you want me to (in that case just shoot me a ticket: https://github.com/infinispan/infinispan-openshift-templates/issues). And let me say it one more time - I'm open for suggestions (and pull requests) if you think this is not the way it should be done.

I don't know how the overarching OpenShift caching, or shared memory services will be exposed, as an OpenShift user that was to store data in Infinispan, I should be able to provide how much (total) data I will put on it, and optionally how many backups I want for the data, and OpenShift should maybe provide with some options on how to do this: 

User: I want 2gb of data
OpenShift: Assuming default of 1 backup (2 copies of data), I can offer you (assuming at least 25% overhead):

a) 2 nodes of 2b
b) 4 nodes of 1gb
c) 8 nodes of 512mb

And user decides... 

Assuming those higher level OpenShift services consume the Infinispan OpenShift templates, and you try to implement a situation like above, where the user specifies total amount of data, and you decide what options to offer them..., then the template would need to expose number of instances (done already) and memory for each of those instance (not there yet).

Still, I'll try to see if I can get my use case working with only 512mb per node, and use the number of instances as a way to add more memory. However, I feel that only exposing number of instances is not enough...

Btw, this is something that needs to be agreed on and should be part of our Infinispan OpenShift integration specification/plan.

Cheers,

> 
> [1] https://github.com/infinispan/infinispan-openshift-templates/blob/master/templates/infinispan-ephemeral.json#L382
>  
> 
> (Don't reply with: make your data smaller)
> 
> Cheers,
> 
> >
> >
> > Thanks,
> > Sanne
> >
> > >
> > > Thanks,
> > > Sebastian
> > >
> > > [1]
> > > https://github.com/jboss-dockerfiles/infinispan/tree/master/server#adjusting-memory
> > > [2]
> > > https://github.com/jboss-dockerfiles/infinispan/blob/master/server/docker-entrypoint.sh#L303-L308
> > > [3] https://www.youtube.com/watch?v=nWGkvrIPqJ4
> > > [4]
> > > https://docs.openshift.com/enterprise/3.2/dev_guide/compute_resources.html
> > >
> > > On Fri, Sep 22, 2017 at 2:33 PM Galder Zamarreño <galder at redhat.com> wrote:
> > >>
> > >> Hi Sebastian,
> > >>
> > >> How do you change memory settings for Infinispan started via service
> > >> catalog?
> > >>
> > >> The memory settings seem defined in [1], but this is not one of the
> > >> parameters supported.
> > >>
> > >> I guess we want this as parameter?
> > >>
> > >> Cheers,
> > >>
> > >> [1]
> > >> https://github.com/infinispan/infinispan-openshift-templates/blob/master/templates/infinispan-ephemeral.json#L308
> > >> --
> > >> Galder Zamarreño
> > >> Infinispan, Red Hat
> > >>
> > >
> > > _______________________________________________
> > > infinispan-dev mailing list
> > > infinispan-dev at lists.jboss.org
> > > https://lists.jboss.org/mailman/listinfo/infinispan-dev
> >
> > _______________________________________________
> > infinispan-dev mailing list
> > infinispan-dev at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/infinispan-dev
> > _______________________________________________
> > infinispan-dev mailing list
> > infinispan-dev at lists.jboss.org
> > https://lists.jboss.org/mailman/listinfo/infinispan-dev
> 
> --
> Galder Zamarreño
> Infinispan, Red Hat
> 
> 
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev
> _______________________________________________
> infinispan-dev mailing list
> infinispan-dev at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/infinispan-dev

--
Galder Zamarreño
Infinispan, Red Hat